includes/Parser.php

   1 <?php
   2
   3 require_once('Tokenizer.php');
   4
   5 if( $GLOBALS['wgUseWikiHiero'] ){
   6         require_once('extensions/wikihiero/wikihiero.php');
   7 }
   8 if( $GLOBALS['wgUseTimeline'] ){
   9         require_once('extensions/timeline/Timeline.php');
  10 }
  11
  12 # PHP Parser
  13 #
  14 # Processes wiki markup
  15 #
  16 # There are two main entry points into the Parser class: parse() and preSaveTransform().
  17 # The parse() function produces HTML output, preSaveTransform() produces altered wiki markup.
  18 #
  19 # Globals used:
  20 #    objects:   $wgLang, $wgDateFormatter, $wgLinkCache, $wgCurParser
  21 #
  22 # NOT $wgArticle, $wgUser or $wgTitle. Keep them away!
  23 #
  24 #    settings:  $wgUseTex*, $wgUseCategoryMagic*, $wgUseDynamicDates*, $wgInterwikiMagic*,
  25 #               $wgNamespacesWithSubpages, $wgLanguageCode, $wgAllowExternalImages*,
  26 #               $wgLocaltimezone
  27 #
  28 #      * only within ParserOptions
  29 #
  30 #
  31 #----------------------------------------
  32 #    Variable substitution O(N^2) attack
  33 #-----------------------------------------
  34 # Without countermeasures, it would be possible to attack the parser by saving a page
  35 # filled with a large number of inclusions of large pages. The size of the generated
  36 # page would be proportional to the square of the input size. Hence, we limit the number
  37 # of inclusions of any given page, thus bringing any attack back to O(N).
  38 #
  39
  40 define( "MAX_INCLUDE_REPEAT", 5 );
  41
  42 # Allowed values for $mOutputType
  43 define( "OT_HTML", 1 );
  44 define( "OT_WIKI", 2 );
  45 define( "OT_MSG", 3 );
  46
  47 # string parameter for extractTags which will cause it
  48 # to strip HTML comments in addition to regular
  49 # <XML>-style tags. This should not be anything we
  50 # may want to use in wikisyntax
  51 define( "STRIP_COMMENTS", "HTMLCommentStrip" );
  52
  53 # prefix for escaping, used in two functions at least
  54 define( "UNIQ_PREFIX", "NaodW29");
  55
  56 class Parser
  57 {
  58         # Cleared with clearState():
  59         var $mOutput, $mAutonumber, $mDTopen, $mStripState = array();
  60         var $mVariables, $mIncludeCount, $mArgStack, $mLastSection, $mInPre;
  61
  62         # Temporary:
  63         var $mOptions, $mTitle, $mOutputType;
  64
  65         function Parser()
  66         {
  67                 $this->clearState();
  68         }
  69
  70         function clearState()
  71         {
  72                 $this->mOutput = new ParserOutput;
  73                 $this->mAutonumber = 0;
  74                 $this->mLastSection = "";
  75                 $this->mDTopen = false;
  76                 $this->mVariables = false;
  77                 $this->mIncludeCount = array();
  78                 $this->mStripState = array();
  79                 $this->mArgStack = array();
  80         }
  81
  82         # First pass--just handle <nowiki> sections, pass the rest off
  83         # to internalParse() which does all the real work.
  84         #
  85         # Returns a ParserOutput
  86         #
  87         function parse( $text, &$title, $options, $linestart = true, $clearState = true )
  88         {
  89                 global $wgUseTidy;
  90                 $fname = "Parser::parse";
  91                 wfProfileIn( $fname );
  92
  93                 if ( $clearState ) {
  94                         $this->clearState();
  95                 }
  96
  97                 $this->mOptions = $options;
  98                 $this->mTitle =& $title;
  99                 $this->mOutputType = OT_HTML;
 100
 101                 $stripState = NULL;
 102                 $text = $this->strip( $text, $this->mStripState );
 103                 $text = $this->internalParse( $text, $linestart );
 104                 $text = $this->unstrip( $text, $this->mStripState );
 105                 # Clean up special characters, only run once, next-to-last before doBlockLevels
 106                 if(!$wgUseTidy) {
 107                         $fixtags = array(
 108                                 "/<hr *>/i" => '<hr/>',
 109                                 "/<br *>/i" => '<br/>',
 110                                 "/<center *>/i"=>'<div class="center">',
 111                                 "/<\\/center *>/i" => '</div>',
 112                                 # Clean up spare ampersands; note that we probably ought to be
 113                                 # more careful about named entities.
 114                                 '/&(?!:amp;|#[Xx][0-9A-fa-f]+;|#[0-9]+;|[a-zA-Z0-9]+;)/' => '&amp;'
 115                         );
 116                         $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 117                 } else {
 118                         $fixtags = array(
 119                                 "/<center *>/i"=>'<div class="center">',
 120                                 "/<\\/center *>/i" => '</div>'
 121                         );
 122                         $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 123                 }
 124                 # only once and last
 125                 $text = $this->doBlockLevels( $text, $linestart );
 126                 if($wgUseTidy) {
 127                         $text = $this->tidy($text);
 128                 }
 129                 $this->mOutput->setText( $text );
 130                 wfProfileOut( $fname );
 131                 return $this->mOutput;
 132         }
 133
 134         /* static */ function getRandomString()
 135         {
 136                 return dechex(mt_rand(0, 0x7fffffff)) . dechex(mt_rand(0, 0x7fffffff));
 137         }
 138
 139         # Replaces all occurrences of <$tag>content</$tag> in the text
 140         # with a random marker and returns the new text. the output parameter
 141         # $content will be an associative array filled with data on the form
 142         # $unique_marker => content.
 143
 144         # If $content is already set, the additional entries will be appended
 145
 146         # If $tag is set to STRIP_COMMENTS, the function will extract
 147         # <!-- HTML comments -->
 148
 149         /* static */ function extractTags($tag, $text, &$content, $uniq_prefix = ""){
 150                 $rnd = $uniq_prefix . '-' . $tag . Parser::getRandomString();
 151                 if ( !$content ) {
 152                         $content = array( );
 153                 }
 154                 $n = 1;
 155                 $stripped = "";
 156
 157                 while ( "" != $text ) {
 158                         if($tag==STRIP_COMMENTS) {
 159                                 $p = preg_split( "/<!--/i", $text, 2 );
 160                         } else {
 161                                 $p = preg_split( "/<\\s*$tag\\s*>/i", $text, 2 );
 162                         }
 163                         $stripped .= $p[0];
 164                         if ( ( count( $p ) < 2 ) || ( "" == $p[1] ) ) {
 165                                 $text = "";
 166                         } else {
 167                                 if($tag==STRIP_COMMENTS) {
 168                                         $q = preg_split( "/-->/i", $p[1], 2 );
 169                                 } else {
 170                                         $q = preg_split( "/<\\/\\s*$tag\\s*>/i", $p[1], 2 );
 171                                 }
 172                                 $marker = $rnd . sprintf("%08X", $n++);
 173                                 $content[$marker] = $q[0];
 174                                 $stripped .= $marker;
 175                                 $text = $q[1];
 176                         }
 177                 }
 178                 return $stripped;
 179         }
 180
 181         # Strips and renders <nowiki>, <pre>, <math>, <hiero>
 182         # If $render is set, performs necessary rendering operations on plugins
 183         # Returns the text, and fills an array with data needed in unstrip()
 184         # If the $state is already a valid strip state, it adds to the state
 185
 186         # When $stripcomments is set, HTML comments <!-- like this -->
 187         # will be stripped in addition to other tags. This is important
 188         # for section editing, where these comments cause confusion when
 189         # counting the sections in the wikisource
 190         function strip( $text, &$state, $stripcomments = false )
 191         {
 192                 $render = ($this->mOutputType == OT_HTML);
 193                 $nowiki_content = array();
 194                 $hiero_content = array();
 195                 $math_content = array();
 196                 $pre_content = array();
 197                 $comment_content = array();
 198
 199                 # Replace any instances of the placeholders
 200                 $uniq_prefix = UNIQ_PREFIX;
 201                 #$text = str_replace( $uniq_prefix, wfHtmlEscapeFirst( $uniq_prefix ), $text );
 202
 203                 $text = Parser::extractTags("nowiki", $text, $nowiki_content, $uniq_prefix);
 204                 foreach( $nowiki_content as $marker => $content ){
 205                         if( $render ){
 206                                 $nowiki_content[$marker] = wfEscapeHTMLTagsOnly( $content );
 207                         } else {
 208                                 $nowiki_content[$marker] = "<nowiki>$content</nowiki>";
 209                         }
 210                 }
 211
 212                 $text = Parser::extractTags("hiero", $text, $hiero_content, $uniq_prefix);
 213                 foreach( $hiero_content as $marker => $content ){
 214                         if( $render && $GLOBALS['wgUseWikiHiero']){
 215                                 $hiero_content[$marker] = WikiHiero( $content, WH_MODE_HTML);
 216                         } else {
 217                                 $hiero_content[$marker] = "<hiero>$content</hiero>";
 218                         }
 219                 }
 220
 221                 $text = Parser::extractTags("math", $text, $math_content, $uniq_prefix);
 222                 foreach( $math_content as $marker => $content ){
 223                         if( $render ) {
 224                                 if( $this->mOptions->getUseTeX() ) {
 225                                         $math_content[$marker] = renderMath( $content );
 226                                 } else {
 227                                         $math_content[$marker] = "&lt;math&gt;$content&lt;math&gt;";
 228                                 }
 229                         } else {
 230                                 $math_content[$marker] = "<math>$content</math>";
 231                         }
 232                 }
 233
 234                 $text = Parser::extractTags("pre", $text, $pre_content, $uniq_prefix);
 235                 foreach( $pre_content as $marker => $content ){
 236                         if( $render ){
 237                                 $pre_content[$marker] = "<pre>" . wfEscapeHTMLTagsOnly( $content ) . "</pre>";
 238                         } else {
 239                                 $pre_content[$marker] = "<pre>$content</pre>";
 240                         }
 241                 }
 242                 if($stripcomments) {
 243                         $text = Parser::extractTags(STRIP_COMMENTS, $text, $comment_content, $uniq_prefix);
 244                         foreach( $comment_content as $marker => $content ){
 245                                 $comment_content[$marker] = "<!--$content-->";
 246                         }
 247                 }
 248
 249                 # Merge state with the pre-existing state, if there is one
 250                 if ( $state ) {
 251                         $state['nowiki'] = $state['nowiki'] + $nowiki_content;
 252                         $state['hiero'] = $state['hiero'] + $hiero_content;
 253                         $state['math'] = $state['math'] + $math_content;
 254                         $state['pre'] = $state['pre'] + $pre_content;
 255                         $state['comment'] = $state['comment'] + $comment_content;
 256                 } else {
 257                         $state = array(
 258                           'nowiki' => $nowiki_content,
 259                           'hiero' => $hiero_content,
 260                           'math' => $math_content,
 261                           'pre' => $pre_content,
 262                           'comment' => $comment_content
 263                         );
 264                 }
 265                 return $text;
 266         }
 267
 268         function unstrip( $text, &$state )
 269         {
 270                 # Must expand in reverse order, otherwise nested tags will be corrupted
 271                 $contentDict = end( $state );
 272                 for ( $contentDict = end( $state ); $contentDict !== false; $contentDict = prev( $state ) ) {
 273                         for ( $content = end( $contentDict ); $content !== false; $content = prev( $contentDict ) ) {
 274                                 $text = str_replace( key( $contentDict ), $content, $text );
 275                         }
 276                 }
 277
 278                 return $text;
 279         }
 280
 281         # Add an item to the strip state
 282         # Returns the unique tag which must be inserted into the stripped text
 283         # The tag will be replaced with the original text in unstrip()
 284
 285         function insertStripItem( $text, &$state )
 286         {
 287                 $rnd = UNIQ_PREFIX . '-item' . Parser::getRandomString();
 288                 if ( !$state ) {
 289                         $state = array(
 290                           'nowiki' => array(),
 291                           'hiero' => array(),
 292                           'math' => array(),
 293                           'pre' => array()
 294                         );
 295                 }
 296                 $state['item'][$rnd] = $text;
 297                 return $rnd;
 298         }
 299
 300         # This method generates the list of subcategories and pages for a category
 301         function categoryMagic ()
 302         {
 303                 global $wgLang , $wgUser ;
 304                 if ( !$this->mOptions->getUseCategoryMagic() ) return ; # Doesn't use categories at all
 305
 306                 $cns = Namespace::getCategory() ;
 307                 if ( $this->mTitle->getNamespace() != $cns ) return "" ; # This ain't a category page
 308
 309                 $r = "<br style=\"clear:both;\"/>\n";
 310
 311
 312                 $sk =& $wgUser->getSkin() ;
 313
 314                 $articles = array() ;
 315                 $children = array() ;
 316                 $data = array () ;
 317                 $id = $this->mTitle->getArticleID() ;
 318
 319                 # For existing categories
 320                 if( $id ) {
 321                         $sql = "SELECT DISTINCT cur_title,cur_namespace FROM cur,links WHERE l_to={$id} AND l_from=cur_id";
 322                         $res = wfQuery ( $sql, DB_READ ) ;
 323                         while ( $x = wfFetchObject ( $res ) ) $data[] = $x ;
 324                 } else {
 325                         # For non-existing categories
 326                         $t = wfStrencode( $this->mTitle->getPrefixedDBKey() );
 327                         $sql = "SELECT DISTINCT cur_title,cur_namespace FROM cur,brokenlinks WHERE bl_to='$t' AND bl_from=cur_id" ;
 328                         $res = wfQuery ( $sql, DB_READ ) ;
 329                         while ( $x = wfFetchObject ( $res ) ) $data[] = $x ;
 330                 }
 331
 332                 # For all pages that link to this category
 333                 foreach ( $data AS $x )
 334                 {
 335                         $t = $wgLang->getNsText ( $x->cur_namespace ) ;
 336                         if ( $t != "" ) $t .= ":" ;
 337                         $t .= $x->cur_title ;
 338
 339                         if ( $x->cur_namespace == $cns ) {
 340                                 array_push ( $children , $sk->makeLink ( $t ) ) ; # Subcategory
 341                         } else {
 342                                 array_push ( $articles , $sk->makeLink ( $t ) ) ; # Page in this category
 343                         }
 344                 }
 345                 wfFreeResult ( $res ) ;
 346
 347                 # Showing subcategories
 348                 if ( count ( $children ) > 0 )
 349                 {
 350                         asort ( $children ) ;
 351                         $r .= "<h2>".wfMsg("subcategories")."</h2>\n" ;
 352                         $r .= implode ( ", " , $children ) ;
 353                 }
 354
 355                 # Showing pages in this category
 356                 if ( count ( $articles ) > 0 )
 357                 {
 358                         $ti = $this->mTitle->getText() ;
 359                         asort ( $articles ) ;
 360                         $h =  wfMsg( "category_header", $ti );
 361                         $r .= "<h2>{$h}</h2>\n" ;
 362                         $r .= implode ( ", " , $articles ) ;
 363                 }
 364
 365
 366                 return $r ;
 367         }
 368
 369         function getHTMLattrs ()
 370         {
 371                 $htmlattrs = array( # Allowed attributes--no scripting, etc.
 372                                 "title", "align", "lang", "dir", "width", "height",
 373                                 "bgcolor", "clear", /* BR */ "noshade", /* HR */
 374                                 "cite", /* BLOCKQUOTE, Q */ "size", "face", "color",
 375                                 /* FONT */ "type", "start", "value", "compact",
 376                                 /* For various lists, mostly deprecated but safe */
 377                                 "summary", "width", "border", "frame", "rules",
 378                                 "cellspacing", "cellpadding", "valign", "char",
 379                                 "charoff", "colgroup", "col", "span", "abbr", "axis",
 380                                 "headers", "scope", "rowspan", "colspan", /* Tables */
 381                                 "id", "class", "name", "style" /* For CSS */
 382                                 );
 383                 return $htmlattrs ;
 384         }
 385
 386         function fixTagAttributes ( $t )
 387         {
 388                 if ( trim ( $t ) == "" ) return "" ; # Saves runtime ;-)
 389                 $htmlattrs = $this->getHTMLattrs() ;
 390
 391                 # Strip non-approved attributes from the tag
 392                 $t = preg_replace(
 393                         "/(\\w+)(\\s*=\\s*([^\\s\">]+|\"[^\">]*\"))?/e",
 394                         "(in_array(strtolower(\"\$1\"),\$htmlattrs)?(\"\$1\".((\"x\$3\" != \"x\")?\"=\$3\":'')):'')",
 395                         $t);
 396                 # Strip javascript "expression" from stylesheets. Brute force approach:
 397                 # If anythin offensive is found, all attributes of the HTML tag are dropped
 398
 399                 if( preg_match(
 400                         "/style\\s*=.*(expression|tps*:\/\/|url\\s*\().*/is",
 401                         wfMungeToUtf8( $t ) ) )
 402                 {
 403                         $t="";
 404                 }
 405
 406                 return trim ( $t ) ;
 407         }
 408
 409         /* interface with html tidy, used if $wgUseTidy = true */
 410         function tidy ( $text ) {
 411                 global $wgTidyConf, $wgTidyBin, $wgTidyOpts;
 412                 global $wgInputEncoding, $wgOutputEncoding;
 413                 $cleansource = '';
 414                 switch(strtoupper($wgOutputEncoding)) {
 415                         case 'ISO-8859-1':
 416                                 $wgTidyOpts .= ($wgInputEncoding == $wgOutputEncoding)? ' -latin1':' -raw';
 417                                 break;
 418                         case 'UTF-8':
 419                                 $wgTidyOpts .= ($wgInputEncoding == $wgOutputEncoding)? ' -utf8':' -raw';
 420                                 break;
 421                         default:
 422                                 $wgTidyOpts .= ' -raw';
 423                         }
 424
 425                 $text = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"'.
 426 ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html>'.
 427 '<head><title>test</title></head><body>'.$text.'</body></html>';
 428                 $descriptorspec = array(
 429                         0 => array("pipe", "r"),
 430                         1 => array("pipe", "w"),
 431                         2 => array("file", "/dev/null", "a")
 432                 );
 433                 $process = proc_open("$wgTidyBin -config $wgTidyConf $wgTidyOpts", $descriptorspec, $pipes);
 434                 if (is_resource($process)) {
 435                         fwrite($pipes[0], $text);
 436                         fclose($pipes[0]);
 437                         while (!feof($pipes[1])) {
 438                                 $cleansource .= fgets($pipes[1], 1024);
 439                         }
 440                         fclose($pipes[1]);
 441                         $return_value = proc_close($process);
 442                 }
 443                 if( $cleansource == '' && $text != '') {
 444                         return '<h2>'.wfMsg('seriousxhtmlerrors').'</h2><pre>'.htmlspecialchars($text).'</pre>';
 445                 } else {
 446                         return $cleansource;
 447                 }
 448         }
 449
 450         function doTableStuff ( $t )
 451         {
 452                 $t = explode ( "\n" , $t ) ;
 453                 $td = array () ; # Is currently a td tag open?
 454                         $ltd = array () ; # Was it TD or TH?
 455                         $tr = array () ; # Is currently a tr tag open?
 456                         $ltr = array () ; # tr attributes
 457                         foreach ( $t AS $k => $x )
 458                         {
 459                                 $x = trim ( $x ) ;
 460                                 $fc = substr ( $x , 0 , 1 ) ;
 461                                 if ( "{|" == substr ( $x , 0 , 2 ) )
 462                                 {
 463                                         $t[$k] = "\n<table " . $this->fixTagAttributes ( substr ( $x , 3 ) ) . ">" ;
 464                                         array_push ( $td , false ) ;
 465                                         array_push ( $ltd , "" ) ;
 466                                         array_push ( $tr , false ) ;
 467                                         array_push ( $ltr , "" ) ;
 468                                 }
 469                                 else if ( count ( $td ) == 0 ) { } # Don't do any of the following
 470                                 else if ( "|}" == substr ( $x , 0 , 2 ) )
 471                                 {
 472                                         $z = "</table>\n" ;
 473                                         $l = array_pop ( $ltd ) ;
 474                                         if ( array_pop ( $tr ) ) $z = "</tr>" . $z ;
 475                                         if ( array_pop ( $td ) ) $z = "</{$l}>" . $z ;
 476                                         array_pop ( $ltr ) ;
 477                                         $t[$k] = $z ;
 478                                 }
 479                                 /*      else if ( "|_" == substr ( $x , 0 , 2 ) ) # Caption
 480                                                 {
 481                                                 $z = trim ( substr ( $x , 2 ) ) ;
 482                                                 $t[$k] = "<caption>{$z}</caption>\n" ;
 483                                                 }*/
 484                                 else if ( "|-" == substr ( $x , 0 , 2 ) ) # Allows for |---------------
 485                                 {
 486                                         $x = substr ( $x , 1 ) ;
 487                                         while ( $x != "" && substr ( $x , 0 , 1 ) == '-' ) $x = substr ( $x , 1 ) ;
 488                                         $z = "" ;
 489                                         $l = array_pop ( $ltd ) ;
 490                                         if ( array_pop ( $tr ) ) $z = "</tr>" . $z ;
 491                                         if ( array_pop ( $td ) ) $z = "</{$l}>" . $z ;
 492                                         array_pop ( $ltr ) ;
 493                                         $t[$k] = $z ;
 494                                         array_push ( $tr , false ) ;
 495                                         array_push ( $td , false ) ;
 496                                         array_push ( $ltd , "" ) ;
 497                                         array_push ( $ltr , $this->fixTagAttributes ( $x ) ) ;
 498                                 }
 499                                 else if ( "|" == $fc || "!" == $fc || "|+" == substr ( $x , 0 , 2 ) ) # Caption
 500                                 {
 501                                         if ( "|+" == substr ( $x , 0 , 2 ) )
 502                                         {
 503                                                 $fc = "+" ;
 504                                                 $x = substr ( $x , 1 ) ;
 505                                         }
 506                                         $after = substr ( $x , 1 ) ;
 507                                         if ( $fc == "!" ) $after = str_replace ( "!!" , "||" , $after ) ;
 508                                         $after = explode ( "||" , $after ) ;
 509                                         $t[$k] = "" ;
 510                                         foreach ( $after AS $theline )
 511                                         {
 512                                                 $z = "" ;
 513                                                 if ( $fc != "+" )
 514                                                 {
 515                                                         $tra = array_pop ( $ltr ) ;
 516                                                         if ( !array_pop ( $tr ) ) $z = "<tr {$tra}>\n" ;
 517                                                         array_push ( $tr , true ) ;
 518                                                         array_push ( $ltr , "" ) ;
 519                                                 }
 520
 521                                                 $l = array_pop ( $ltd ) ;
 522                                                 if ( array_pop ( $td ) ) $z = "</{$l}>" . $z ;
 523                                                 if ( $fc == "|" ) $l = "td" ;
 524                                                 else if ( $fc == "!" ) $l = "th" ;
 525                                                 else if ( $fc == "+" ) $l = "caption" ;
 526                                                 else $l = "" ;
 527                                                 array_push ( $ltd , $l ) ;
 528                                                 $y = explode ( "|" , $theline , 2 ) ;
 529                                                 if ( count ( $y ) == 1 ) $y = "{$z}<{$l}>{$y[0]}" ;
 530                                                 else $y = $y = "{$z}<{$l} ".$this->fixTagAttributes($y[0]).">{$y[1]}" ;
 531                                                 $t[$k] .= $y ;
 532                                                 array_push ( $td , true ) ;
 533                                         }
 534                                 }
 535                         }
 536
 537                 # Closing open td, tr && table
 538                 while ( count ( $td ) > 0 )
 539                 {
 540                         if ( array_pop ( $td ) ) $t[] = "</td>" ;
 541                         if ( array_pop ( $tr ) ) $t[] = "</tr>" ;
 542                         $t[] = "</table>" ;
 543                 }
 544
 545                 $t = implode ( "\n" , $t ) ;
 546                 #               $t = $this->removeHTMLtags( $t );
 547                 return $t ;
 548         }
 549
 550         function internalParse( $text, $linestart, $args = array() )
 551         {
 552                 $fname = "Parser::internalParse";
 553                 wfProfileIn( $fname );
 554
 555                 $text = $this->removeHTMLtags( $text );
 556                 $text = $this->replaceVariables( $text, $args );
 557
 558                 # $text = preg_replace( "/(^|\n)-----*/", "\\1<hr>", $text );
 559
 560                 $text = $this->doHeadings( $text );
 561                 if($this->mOptions->getUseDynamicDates()) {
 562                         global $wgDateFormatter;
 563                         $text = $wgDateFormatter->reformat( $this->mOptions->getDateFormat(), $text );
 564                 }
 565                 $text = $this->replaceExternalLinks( $text );
 566                 $text = $this->doTokenizedParser ( $text );
 567                 $text = $this->doTableStuff ( $text ) ;
 568                 $text = $this->formatHeadings( $text );
 569                 $sk =& $this->mOptions->getSkin();
 570                 $text = $sk->transformContent( $text );
 571
 572                 if ( !isset ( $this->categoryMagicDone ) ) {
 573                    $text .= $this->categoryMagic () ;
 574                    $this->categoryMagicDone = true ;
 575                    }
 576
 577                 wfProfileOut( $fname );
 578                 return $text;
 579         }
 580
 581
 582         /* private */ function doHeadings( $text )
 583         {
 584                 for ( $i = 6; $i >= 1; --$i ) {
 585                         $h = substr( "======", 0, $i );
 586                         $text = preg_replace( "/^{$h}(.+){$h}(\\s|$)/m",
 587                           "<h{$i}>\\1</h{$i}>\\2", $text );
 588                 }
 589                 return $text;
 590         }
 591
 592         # Note: we have to do external links before the internal ones,
 593         # and otherwise take great care in the order of things here, so
 594         # that we don't end up interpreting some URLs twice.
 595
 596         /* private */ function replaceExternalLinks( $text )
 597         {
 598                 $fname = "Parser::replaceExternalLinks";
 599                 wfProfileIn( $fname );
 600                 $text = $this->subReplaceExternalLinks( $text, "http", true );
 601                 $text = $this->subReplaceExternalLinks( $text, "https", true );
 602                 $text = $this->subReplaceExternalLinks( $text, "ftp", false );
 603                 $text = $this->subReplaceExternalLinks( $text, "irc", false );
 604                 $text = $this->subReplaceExternalLinks( $text, "gopher", false );
 605                 $text = $this->subReplaceExternalLinks( $text, "news", false );
 606                 $text = $this->subReplaceExternalLinks( $text, "mailto", false );
 607                 wfProfileOut( $fname );
 608                 return $text;
 609         }
 610
 611         /* private */ function subReplaceExternalLinks( $s, $protocol, $autonumber )
 612         {
 613                 $unique = "4jzAfzB8hNvf4sqyO9Edd8pSmk9rE2in0Tgw3";
 614                 $uc = "A-Za-z0-9_\\/~%\\-+&*#?!=()@\\x80-\\xFF";
 615
 616                 # this is  the list of separators that should be ignored if they
 617                 # are the last character of an URL but that should be included
 618                 # if they occur within the URL, e.g. "go to www.foo.com, where .."
 619                 # in this case, the last comma should not become part of the URL,
 620                 # but in "www.foo.com/123,2342,32.htm" it should.
 621                 $sep = ",;\.:";
 622                 $fnc = "A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF";
 623                 $images = "gif|png|jpg|jpeg";
 624
 625                 # PLEASE NOTE: The curly braces { } are not part of the regex,
 626                 # they are interpreted as part of the string (used to tell PHP
 627                 # that the content of the string should be inserted there).
 628                 $e1 = "/(^|[^\\[])({$protocol}:)([{$uc}{$sep}]+)\\/([{$fnc}]+)\\." .
 629                   "((?i){$images})([^{$uc}]|$)/";
 630
 631                 $e2 = "/(^|[^\\[])({$protocol}:)(([".$uc."]|[".$sep."][".$uc."])+)([^". $uc . $sep. "]|[".$sep."]|$)/";
 632                 $sk =& $this->mOptions->getSkin();
 633
 634                 if ( $autonumber and $this->mOptions->getAllowExternalImages() ) { # Use img tags only for HTTP urls
 635                         $s = preg_replace( $e1, "\\1" . $sk->makeImage( "{$unique}:\\3" .
 636                           "/\\4.\\5", "\\4.\\5" ) . "\\6", $s );
 637                 }
 638                 $s = preg_replace( $e2, "\\1" . "<a href=\"{$unique}:\\3\"" .
 639                   $sk->getExternalLinkAttributes( "{$unique}:\\3", wfEscapeHTML(
 640                   "{$unique}:\\3" ) ) . ">" . wfEscapeHTML( "{$unique}:\\3" ) .
 641                   "</a>\\5", $s );
 642                 $s = str_replace( $unique, $protocol, $s );
 643
 644                 $a = explode( "[{$protocol}:", " " . $s );
 645                 $s = array_shift( $a );
 646                 $s = substr( $s, 1 );
 647
 648                 $e1 = "/^([{$uc}"."{$sep}]+)](.*)\$/sD";
 649                 $e2 = "/^([{$uc}"."{$sep}]+)\\s+([^\\]]+)](.*)\$/sD";
 650
 651                 foreach ( $a as $line ) {
 652                         if ( preg_match( $e1, $line, $m ) ) {
 653                                 $link = "{$protocol}:{$m[1]}";
 654                                 $trail = $m[2];
 655                                 if ( $autonumber ) { $text = "[" . ++$this->mAutonumber . "]"; }
 656                                 else { $text = wfEscapeHTML( $link ); }
 657                         } else if ( preg_match( $e2, $line, $m ) ) {
 658                                 $link = "{$protocol}:{$m[1]}";
 659                                 $text = $m[2];
 660                                 $trail = $m[3];
 661                         } else {
 662                                 $s .= "[{$protocol}:" . $line;
 663                                 continue;
 664                         }
 665                         if( $link == $text || preg_match( "!$protocol://" . preg_quote( $text, "/" ) . "/?$!", $link ) ) {
 666                                 $paren = "";
 667                         } else {
 668                                 # Expand the URL for printable version
 669                                 $paren = "<span class='urlexpansion'> (<i>" . htmlspecialchars ( $link ) . "</i>)</span>";
 670                         }
 671                         $la = $sk->getExternalLinkAttributes( $link, $text );
 672                         $s .= "<a href='{$link}'{$la}>{$text}</a>{$paren}{$trail}";
 673
 674                 }
 675                 return $s;
 676         }
 677
 678         /* private */ function handle3Quotes( &$state, $token )
 679         {
 680                 if ( $state["strong"] !== false ) {
 681                         if ( $state["em"] !== false && $state["em"] > $state["strong"] )
 682                         {
 683                                 # ''' lala ''lala '''
 684                                 $s = "</em></strong><em>";
 685                         } else {
 686                                 $s = "</strong>";
 687                         }
 688                         $state["strong"] = FALSE;
 689                 } else {
 690                         $s = "<strong>";
 691                         $state["strong"] = isset($token["pos"]) ? $token["pos"] : true;
 692                 }
 693                 return $s;
 694         }
 695
 696         /* private */ function handle2Quotes( &$state, $token )
 697         {
 698                 if ( $state["em"] !== false ) {
 699                         if ( $state["strong"] !== false && $state["strong"] > $state["em"] )
 700                         {
 701                                 # ''lala'''lala'' ....'''
 702                                 $s = "</strong></em><strong>";
 703                         } else {
 704                                 $s = "</em>";
 705                         }
 706                         $state["em"] = FALSE;
 707                 } else {
 708                         $s = "<em>";
 709                         $state["em"] = isset($token["pos"]) ? $token["pos"] : true;
 710
 711                 }
 712                 return $s;
 713         }
 714
 715         /* private */ function handle5Quotes( &$state, $token )
 716         {
 717                 $s = "";
 718                 if ( $state["em"] !== false && $state["strong"] !== false ) {
 719                         if ( $state["em"] < $state["strong"] ) {
 720                                 $s .= "</strong></em>";
 721                         } else {
 722                                 $s .= "</em></strong>";
 723                         }
 724                         $state["strong"] = $state["em"] = FALSE;
 725                 } elseif ( $state["em"] !== false ) {
 726                         $s .= "</em><strong>";
 727                         $state["em"] = FALSE;
 728                         $state["strong"] = $token["pos"];
 729                 } elseif ( $state["strong"] !== false ) {
 730                         $s .= "</strong><em>";
 731                         $state["strong"] = FALSE;
 732                         $state["em"] = $token["pos"];
 733                 } else { # not $em and not $strong
 734                         $s .= "<strong><em>";
 735                         $state["strong"] = $state["em"] = isset($token["pos"]) ? $token["pos"] : true;
 736                 }
 737                 return $s;
 738         }
 739
 740         /* private */ function doTokenizedParser( $str )
 741         {
 742                 global $wgLang; # for language specific parser hook
 743                 global $wgUploadDirectory, $wgUseTimeline;
 744
 745                 $tokenizer=Tokenizer::newFromString( $str );
 746                 $tokenStack = array();
 747
 748                 $s="";
 749                 $state["em"]      = FALSE;
 750                 $state["strong"]  = FALSE;
 751                 $tagIsOpen = FALSE;
 752                 $threeopen = false;
 753
 754                 # The tokenizer splits the text into tokens and returns them one by one.
 755                 # Every call to the tokenizer returns a new token.
 756                 while ( $token = $tokenizer->nextToken() )
 757                 {
 758                         switch ( $token["type"] )
 759                         {
 760                                 case "text":
 761                                         # simple text with no further markup
 762                                         $txt = $token["text"];
 763                                         break;
 764                                 case "blank":
 765                                         # Text that contains blanks that have to be converted to
 766                                         # non-breakable spaces for French.
 767                                         # U+202F NARROW NO-BREAK SPACE might be a better choice, but
 768                                         # browser support for Unicode spacing is poor.
 769                                         $txt = str_replace( " ", "&nbsp;", $token["text"] );
 770                                         break;
 771                                 case "[[[":
 772                                         # remember the tag opened with 3 [
 773                                         $threeopen = true;
 774                                 case "[[":
 775                                         # link opening tag.
 776                                         # FIXME : Treat orphaned open tags (stack not empty when text is over)
 777                                         $tagIsOpen = TRUE;
 778                                         array_push( $tokenStack, $token );
 779                                         $txt="";
 780                                         break;
 781
 782                                 case "]]]":
 783                                 case "]]":
 784                                         # link close tag.
 785                                         # get text from stack, glue it together, and call the code to handle a
 786                                         # link
 787
 788                                         if ( count( $tokenStack ) == 0 )
 789                                         {
 790                                                 # stack empty. Found a ]] without an opening [[
 791                                                 $txt = "]]";
 792                                         } else {
 793                                                 $linkText = "";
 794                                                 $lastToken = array_pop( $tokenStack );
 795                                                 while ( !(($lastToken["type"] == "[[[") or ($lastToken["type"] == "[[")) )
 796                                                 {
 797                                                         if( !empty( $lastToken["text"] ) ) {
 798                                                                 $linkText = $lastToken["text"] . $linkText;
 799                                                         }
 800                                                         $lastToken = array_pop( $tokenStack );
 801                                                 }
 802
 803                                                 $txt = $linkText ."]]";
 804
 805                                                 if( isset( $lastToken["text"] ) ) {
 806                                                         $prefix = $lastToken["text"];
 807                                                 } else {
 808                                                         $prefix = "";
 809                                                 }
 810                                                 $nextToken = $tokenizer->previewToken();
 811                                                 if ( $nextToken["type"] == "text" )
 812                                                 {
 813                                                         # Preview just looks at it. Now we have to fetch it.
 814                                                         $nextToken = $tokenizer->nextToken();
 815                                                         $txt .= $nextToken["text"];
 816                                                 }
 817                                                 $txt = $this->handleInternalLink( $this->unstrip($txt,$this->mStripState), $prefix );
 818
 819                                                 # did the tag start with 3 [ ?
 820                                                 if($threeopen) {
 821                                                         # show the first as text
 822                                                         $txt = "[".$txt;
 823                                                         $threeopen=false;
 824                                                 }
 825
 826                                         }
 827                                         $tagIsOpen = (count( $tokenStack ) != 0);
 828                                         break;
 829                                 case "----":
 830                                         $txt = "\n<hr />\n";
 831                                         break;
 832                                 case "'''":
 833                                         # This and the three next ones handle quotes
 834                                         $txt = $this->handle3Quotes( $state, $token );
 835                                         break;
 836                                 case "''":
 837                                         $txt = $this->handle2Quotes( $state, $token );
 838                                         break;
 839                                 case "'''''":
 840                                         $txt = $this->handle5Quotes( $state, $token );
 841                                         break;
 842                                 case "":
 843                                         # empty token
 844                                         $txt="";
 845                                         break;
 846                                 case "RFC ":
 847                                         if ( $tagIsOpen ) {
 848                                                 $txt = "RFC ";
 849                                         } else {
 850                                                 $txt = $this->doMagicRFC( $tokenizer );
 851                                         }
 852                                         break;
 853                                 case "ISBN ":
 854                                         if ( $tagIsOpen ) {
 855                                                 $txt = "ISBN ";
 856                                         } else {
 857                                                 $txt = $this->doMagicISBN( $tokenizer );
 858                                         }
 859                                         break;
 860                                 case "<timeline>":
 861                                         if ( $wgUseTimeline &&
 862                                              "" != ( $timelinesrc = $tokenizer->readAllUntil("&lt;/timeline&gt;") ) )
 863                                         {
 864                                                 $txt = renderTimeline( $timelinesrc );
 865                                         } else {
 866                                                 $txt=$token["text"];
 867                                         }
 868                                         break;
 869                                 default:
 870                                         # Call language specific Hook.
 871                                         $txt = $wgLang->processToken( $token, $tokenStack );
 872                                         if ( NULL == $txt ) {
 873                                                 # An unkown token. Highlight.
 874                                                 $txt = "<font color=\"#FF0000\"><b>".$token["type"]."</b></font>";
 875                                                 $txt .= "<font color=\"#FFFF00\"><b>".$token["text"]."</b></font>";
 876                                         }
 877                                         break;
 878                         }
 879                         # If we're parsing the interior of a link, don't append the interior to $s,
 880                         # but push it to the stack so it can be processed when a ]] token is found.
 881                         if ( $tagIsOpen  && $txt != "" ) {
 882                                 $token["type"] = "text";
 883                                 $token["text"] = $txt;
 884                                 array_push( $tokenStack, $token );
 885                         } else {
 886                                 $s .= $txt;
 887                         }
 888                 } #end while
 889                 if ( count( $tokenStack ) != 0 )
 890                 {
 891                         # still objects on stack. opened [[ tag without closing ]] tag.
 892                         $txt = "";
 893                         while ( $lastToken = array_pop( $tokenStack ) )
 894                         {
 895                                 if ( $lastToken["type"] == "text" )
 896                                 {
 897                                         $txt = $lastToken["text"] . $txt;
 898                                 } else {
 899                                         $txt = $lastToken["type"] . $txt;
 900                                 }
 901                         }
 902                         $s .= $txt;
 903                 }
 904                 return $s;
 905         }
 906
 907         /* private */ function handleInternalLink( $line, $prefix )
 908         {
 909                 global $wgLang, $wgLinkCache;
 910                 global $wgNamespacesWithSubpages, $wgLanguageCode;
 911                 static $fname = "Parser::handleInternalLink" ;
 912                 wfProfileIn( $fname );
 913
 914                 wfProfileIn( "$fname-setup" );
 915                 static $tc = FALSE;
 916                 if ( !$tc ) { $tc = Title::legalChars() . "#"; }
 917                 $sk =& $this->mOptions->getSkin();
 918
 919                 # Match a link having the form [[namespace:link|alternate]]trail
 920                 static $e1 = FALSE;
 921                 if ( !$e1 ) { $e1 = "/^([{$tc}]+)(?:\\|([^]]+))?]](.*)\$/sD"; }
 922                 # Match the end of a line for a word that's not followed by whitespace,
 923                 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
 924                 #$e2 = "/^(.*)\\b(\\w+)\$/suD";
 925                 #$e2 = "/^(.*\\s)(\\S+)\$/suD";
 926                 static $e2 = '/^(.*\s)([a-zA-Z\x80-\xff]+)$/sD';
 927
 928
 929                 # Special and Media are pseudo-namespaces; no pages actually exist in them
 930                 static $image = FALSE;
 931                 static $special = FALSE;
 932                 static $media = FALSE;
 933                 static $category = FALSE;
 934                 if ( !$image ) { $image = Namespace::getImage(); }
 935                 if ( !$special ) { $special = Namespace::getSpecial(); }
 936                 if ( !$media ) { $media = Namespace::getMedia(); }
 937                 if ( !$category ) { $category = Namespace::getCategory(); }
 938
 939                 $nottalk = !Namespace::isTalk( $this->mTitle->getNamespace() );
 940
 941                 wfProfileOut( "$fname-setup" );
 942                 $s = "";
 943
 944                 if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
 945                         $text = $m[2];
 946                         $trail = $m[3];
 947                 } else { # Invalid form; output directly
 948                         $s .= $prefix . "[[" . $line ;
 949                         return $s;
 950                 }
 951
 952                 /* Valid link forms:
 953                 Foobar -- normal
 954                 :Foobar -- override special treatment of prefix (images, language links)
 955                 /Foobar -- convert to CurrentPage/Foobar
 956                 /Foobar/ -- convert to CurrentPage/Foobar, strip the initial / from text
 957                 */
 958                 $c = substr($m[1],0,1);
 959                 $noforce = ($c != ":");
 960                 if( $c == "/" ) { # subpage
 961                         if(substr($m[1],-1,1)=="/") {                 # / at end means we don't want the slash to be shown
 962                                 $m[1]=substr($m[1],1,strlen($m[1])-2);
 963                                 $noslash=$m[1];
 964                         } else {
 965                                 $noslash=substr($m[1],1);
 966                         }
 967                         if($wgNamespacesWithSubpages[$this->mTitle->getNamespace()]) { # subpages allowed here
 968                                 $link = $this->mTitle->getPrefixedText(). "/" . trim($noslash);
 969                                 if( "" == $text ) {
 970                                         $text= $m[1];
 971                                 } # this might be changed for ugliness reasons
 972                         } else {
 973                                 $link = $noslash; # no subpage allowed, use standard link
 974                         }
 975                 } elseif( $noforce ) { # no subpage
 976                         $link = $m[1];
 977                 } else {
 978                         $link = substr( $m[1], 1 );
 979                 }
 980                 if( "" == $text )
 981                         $text = $link;
 982
 983                 $nt = Title::newFromText( $link );
 984                 if( !$nt ) {
 985                         $s .= $prefix . "[[" . $line;
 986                         return $s;
 987                 }
 988                 $ns = $nt->getNamespace();
 989                 $iw = $nt->getInterWiki();
 990                 if( $noforce ) {
 991                         if( $iw && $this->mOptions->getInterwikiMagic() && $nottalk && $wgLang->getLanguageName( $iw ) ) {
 992                                 array_push( $this->mOutput->mLanguageLinks, $nt->getPrefixedText() );
 993                                 $s .= $prefix . $trail ;
 994                                 return (trim($s) == '')? '': $s;
 995                         }
 996                         if( $ns == $image ) {
 997                                 $s .= $prefix . $sk->makeImageLinkObj( $nt, $text ) . $trail;
 998                                 $wgLinkCache->addImageLinkObj( $nt );
 999                                 return $s;
1000                         }
1001                         if ( $ns == $category ) {
1002                                 $t = $nt->getText() ;
1003                                 $nnt = Title::newFromText ( Namespace::getCanonicalName($category).":".$t ) ;
1004                                 $t = $sk->makeLinkObj( $nnt, $t, "", "" , $prefix );
1005                                 $this->mOutput->mCategoryLinks[] = $t ;
1006                                 $s .= $prefix . $trail ;
1007                                 return $s ;
1008                         }
1009                 }
1010                 if( ( $nt->getPrefixedText() == $this->mTitle->getPrefixedText() ) &&
1011                     ( strpos( $link, "#" ) == FALSE ) ) {
1012                         # Self-links are handled specially; generally de-link and change to bold.
1013                         $s .= $prefix . $sk->makeSelfLinkObj( $nt, $text, "", $trail );
1014                         return $s;
1015                 }
1016
1017                 if( $ns == $media ) {
1018                         $s .= $prefix . $sk->makeMediaLinkObj( $nt, $text ) . $trail;
1019                         $wgLinkCache->addImageLinkObj( $nt );
1020                         return $s;
1021                 } elseif( $ns == $special ) {
1022                         $s .= $prefix . $sk->makeKnownLinkObj( $nt, $text, "", $trail );
1023                         return $s;
1024                 }
1025                 $s .= $sk->makeLinkObj( $nt, $text, "", $trail , $prefix );
1026
1027                 wfProfileOut( $fname );
1028                 return $s;
1029         }
1030
1031         # Some functions here used by doBlockLevels()
1032         #
1033         /* private */ function closeParagraph()
1034         {
1035                 $result = "";
1036                 if ( '' != $this->mLastSection ) {
1037                         $result = "</" . $this->mLastSection  . ">\n";
1038                 }
1039                 $this->mInPre = false;
1040                 $this->mLastSection = "";
1041                 return $result;
1042         }
1043         # getCommon() returns the length of the longest common substring
1044         # of both arguments, starting at the beginning of both.
1045         #
1046         /* private */ function getCommon( $st1, $st2 )
1047         {
1048                 $fl = strlen( $st1 );
1049                 $shorter = strlen( $st2 );
1050                 if ( $fl < $shorter ) { $shorter = $fl; }
1051
1052                 for ( $i = 0; $i < $shorter; ++$i ) {
1053                         if ( $st1{$i} != $st2{$i} ) { break; }
1054                 }
1055                 return $i;
1056         }
1057         # These next three functions open, continue, and close the list
1058         # element appropriate to the prefix character passed into them.
1059         #
1060         /* private */ function openList( $char )
1061     {
1062                 $result = $this->closeParagraph();
1063
1064                 if ( "*" == $char ) { $result .= "<ul><li>"; }
1065                 else if ( "#" == $char ) { $result .= "<ol><li>"; }
1066                 else if ( ":" == $char ) { $result .= "<dl><dd>"; }
1067                 else if ( ";" == $char ) {
1068                         $result .= "<dl><dt>";
1069                         $this->mDTopen = true;
1070                 }
1071                 else { $result = "<!-- ERR 1 -->"; }
1072
1073                 return $result;
1074         }
1075
1076         /* private */ function nextItem( $char )
1077         {
1078                 if ( "*" == $char || "#" == $char ) { return "</li><li>"; }
1079                 else if ( ":" == $char || ";" == $char ) {
1080                         $close = "</dd>";
1081                         if ( $this->mDTopen ) { $close = "</dt>"; }
1082                         if ( ";" == $char ) {
1083                                 $this->mDTopen = true;
1084                                 return $close . "<dt>";
1085                         } else {
1086                                 $this->mDTopen = false;
1087                                 return $close . "<dd>";
1088                         }
1089                 }
1090                 return "<!-- ERR 2 -->";
1091         }
1092
1093         /* private */function closeList( $char )
1094         {
1095                 if ( "*" == $char ) { $text = "</li></ul>"; }
1096                 else if ( "#" == $char ) { $text = "</li></ol>"; }
1097                 else if ( ":" == $char ) {
1098                         if ( $this->mDTopen ) {
1099                                 $this->mDTopen = false;
1100                                 $text = "</dt></dl>";
1101                         } else {
1102                                 $text = "</dd></dl>";
1103                         }
1104                 }
1105                 else {  return "<!-- ERR 3 -->"; }
1106                 return $text."\n";
1107         }
1108
1109         /* private */ function doBlockLevels( $text, $linestart ) {
1110                 $fname = "Parser::doBlockLevels";
1111                 wfProfileIn( $fname );
1112
1113                 # Parsing through the text line by line.  The main thing
1114                 # happening here is handling of block-level elements p, pre,
1115                 # and making lists from lines starting with * # : etc.
1116                 #
1117                 $textLines = explode( "\n", $text );
1118
1119                 $lastPrefix = $output = $lastLine = '';
1120                 $this->mDTopen = $inBlockElem = false;
1121                 $prefixLength = 0;
1122                 $paragraphStack = false;
1123
1124                 if ( !$linestart ) {
1125                         $output .= array_shift( $textLines );
1126                 }
1127                 foreach ( $textLines as $oLine ) {
1128                         $lastPrefixLength = strlen( $lastPrefix );
1129                         $preCloseMatch = preg_match("/<\\/pre/i", $oLine );
1130                         $preOpenMatch = preg_match("/<pre/i", $oLine );
1131                         if (!$this->mInPre) {
1132                                 $this->mInPre = !empty($preOpenMatch);
1133                         }
1134                         if ( !$this->mInPre ) {
1135                                 # Multiple prefixes may abut each other for nested lists.
1136                                 $prefixLength = strspn( $oLine, "*#:;" );
1137                                 $pref = substr( $oLine, 0, $prefixLength );
1138
1139                                 # eh?
1140                                 $pref2 = str_replace( ";", ":", $pref );
1141                                 $t = substr( $oLine, $prefixLength );
1142                         } else {
1143                                 # Don't interpret any other prefixes in preformatted text
1144                                 $prefixLength = 0;
1145                                 $pref = $pref2 = '';
1146                                 $t = $oLine;
1147                         }
1148
1149                         # List generation
1150                         if( $prefixLength && 0 == strcmp( $lastPrefix, $pref2 ) ) {
1151                                 # Same as the last item, so no need to deal with nesting or opening stuff
1152                                 $output .= $this->nextItem( substr( $pref, -1 ) );
1153                                 $paragraphStack = false;
1154
1155                                 if ( ";" == substr( $pref, -1 ) ) {
1156                                         # The one nasty exception: definition lists work like this:
1157                                         # ; title : definition text
1158                                         # So we check for : in the remainder text to split up the
1159                                         # title and definition, without b0rking links.
1160                                         # FIXME: This is not foolproof. Something better in Tokenizer might help.
1161                                         if( preg_match( '/^(.*?(?:\s|&nbsp;)):(.*)$/', $t, $match ) ) {
1162                                                 $term = $match[1];
1163                                                 $output .= $term . $this->nextItem( ":" );
1164                                                 $t = $match[2];
1165                                         }
1166                                 }
1167                         } elseif( $prefixLength || $lastPrefixLength ) {
1168                                 # Either open or close a level...
1169                                 $commonPrefixLength = $this->getCommon( $pref, $lastPrefix );
1170                                 $paragraphStack = false;
1171
1172                                 while( $commonPrefixLength < $lastPrefixLength ) {
1173                                         $output .= $this->closeList( $lastPrefix{$lastPrefixLength-1} );
1174                                         --$lastPrefixLength;
1175                                 }
1176                                 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
1177                                         $output .= $this->nextItem( $pref{$commonPrefixLength-1} );
1178                                 }
1179                                 while ( $prefixLength > $commonPrefixLength ) {
1180                                         $char = substr( $pref, $commonPrefixLength, 1 );
1181                                         $output .= $this->openList( $char );
1182
1183                                         if ( ";" == $char ) {
1184                                                 # FIXME: This is dupe of code above
1185                                                 if( preg_match( '/^(.*?(?:\s|&nbsp;)):(.*)$/', $t, $match ) ) {
1186                                                         $term = $match[1];
1187                                                         $output .= $term . $this->nextItem( ":" );
1188                                                         $t = $match[2];
1189                                                 }
1190                                         }
1191                                         ++$commonPrefixLength;
1192                                 }
1193                                 $lastPrefix = $pref2;
1194                         }
1195                         if( 0 == $prefixLength ) {
1196                                 # No prefix (not in list)--go to paragraph mode
1197                                 $uniq_prefix = UNIQ_PREFIX;
1198                                 // XXX: use a stack for nestable elements like span, table and div
1199                                 $openmatch = preg_match("/(<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<div|<pre|<tr|<td|<p|<ul|<li)/i", $t );
1200                                 $closematch = preg_match(
1201                                         "/(<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|".
1202                                         "<\\/div|<hr|<\\/td|<\\/pre|<\\/p|".$uniq_prefix."-pre|<\\/li|<\\/ul)/i", $t );
1203                                 if ( $openmatch or $closematch ) {
1204                                         $paragraphStack = false;
1205                                         $output .= $this->closeParagraph();
1206                                         if($preOpenMatch and !$preCloseMatch) {
1207                                                 $this->mInPre = true;
1208                                         }
1209                                         if ( $closematch  ) {
1210                                                 $inBlockElem = false;
1211                                         } else {
1212                                                 $inBlockElem = true;
1213                                         }
1214                                 } else if ( !$inBlockElem ) {
1215                                         if ( " " == $t{0} ) {
1216                                                 // pre
1217                                                 if ($this->mLastSection != 'pre') {
1218                                                         $paragraphStack = false;
1219                                                         $output .= $this->closeParagraph().'<pre>';
1220                                                         $this->mLastSection = 'pre';
1221                                                 }
1222                                         } else {
1223                                                 // paragraph
1224                                                 if ( '' == trim($t) ) {
1225                                                         if ( $paragraphStack ) {
1226                                                                 $output .= $paragraphStack.'<br/>';
1227                                                                 $paragraphStack = false;
1228                                                                 $this->mLastSection = 'p';
1229                                                         } else {
1230                                                                 if ($this->mLastSection != 'p' ) {
1231                                                                         $output .= $this->closeParagraph();
1232                                                                         $this->mLastSection = '';
1233                                                                         $paragraphStack = "<p>";
1234                                                                 } else {
1235                                                                         $paragraphStack = '</p><p>';
1236                                                                 }
1237                                                         }
1238                                                 } else {
1239                                                         if ( $paragraphStack ) {
1240                                                                 $output .= $paragraphStack;
1241                                                                 $paragraphStack = false;
1242                                                                 $this->mLastSection = 'p';
1243                                                         } else if ($this->mLastSection != 'p') {
1244                                                                 $output .= $this->closeParagraph().'<p>';
1245                                                                 $this->mLastSection = 'p';
1246                                                         }
1247                                                 }
1248                                         }
1249                                 }
1250                         }
1251                         if ($paragraphStack === false) {
1252                                 $output .= $t."\n";
1253                         }
1254                 }
1255                 while ( $prefixLength ) {
1256                         $output .= $this->closeList( $pref2{$prefixLength-1} );
1257                         --$prefixLength;
1258                 }
1259                 if ( "" != $this->mLastSection ) {
1260                         $output .= "</" . $this->mLastSection . ">";
1261                         $this->mLastSection = "";
1262                 }
1263
1264                 wfProfileOut( $fname );
1265                 return $output;
1266         }
1267
1268         function getVariableValue( $index ) {
1269                 global $wgLang, $wgSitename, $wgServer;
1270
1271                 switch ( $index ) {
1272                         case MAG_CURRENTMONTH:
1273                                 return date( "m" );
1274                         case MAG_CURRENTMONTHNAME:
1275                                 return $wgLang->getMonthName( date("n") );
1276                         case MAG_CURRENTMONTHNAMEGEN:
1277                                 return $wgLang->getMonthNameGen( date("n") );
1278                         case MAG_CURRENTDAY:
1279                                 return date("j");
1280                         case MAG_PAGENAME:
1281                                 return $this->mTitle->getText();
1282                         case MAG_NAMESPACE:
1283                                 # return Namespace::getCanonicalName($this->mTitle->getNamespace());
1284                                 return $wgLang->getNsText($this->mTitle->getNamespace()); // Patch  by Dori
1285                         case MAG_CURRENTDAYNAME:
1286                                 return $wgLang->getWeekdayName( date("w")+1 );
1287                         case MAG_CURRENTYEAR:
1288                                 return date( "Y" );
1289                         case MAG_CURRENTTIME:
1290                                 return $wgLang->time( wfTimestampNow(), false );
1291                         case MAG_NUMBEROFARTICLES:
1292                                 return wfNumberOfArticles();
1293                         case MAG_SITENAME:
1294                                 return $wgSitename;
1295                         case MAG_SERVER:
1296                                 return $wgServer;
1297                         default:
1298                                 return NULL;
1299                 }
1300         }
1301
1302         function initialiseVariables()
1303         {
1304                 global $wgVariableIDs;
1305                 $this->mVariables = array();
1306                 foreach ( $wgVariableIDs as $id ) {
1307                         $mw =& MagicWord::get( $id );
1308                         $mw->addToArray( $this->mVariables, $this->getVariableValue( $id ) );
1309                 }
1310         }
1311
1312         /* private */ function replaceVariables( $text, $args = array() )
1313         {
1314                 global $wgLang, $wgScript, $wgArticlePath;
1315
1316                 $fname = "Parser::replaceVariables";
1317                 wfProfileIn( $fname );
1318
1319                 $bail = false;
1320                 if ( !$this->mVariables ) {
1321                         $this->initialiseVariables();
1322                 }
1323                 $titleChars = Title::legalChars();
1324                 $regex = "/(\\n?){{([$titleChars]*?)(\\|.*?|)}}/s";
1325
1326                 # This function is called recursively. To keep track of arguments we need a stack:
1327                 array_push( $this->mArgStack, $args );
1328
1329                 # PHP global rebinding syntax is a bit weird, need to use the GLOBALS array
1330                 $GLOBALS['wgCurParser'] =& $this;
1331                 $text = preg_replace_callback( $regex, "wfBraceSubstitution", $text );
1332
1333                 array_pop( $this->mArgStack );
1334
1335                 return $text;
1336         }
1337
1338         function braceSubstitution( $matches )
1339         {
1340                 global $wgLinkCache, $wgLang;
1341                 $fname = "Parser::braceSubstitution";
1342                 $found = false;
1343                 $nowiki = false;
1344                 $title = NULL;
1345
1346                 # $newline is an optional newline character before the braces
1347                 # $part1 is the bit before the first |, and must contain only title characters
1348                 # $args is a list of arguments, starting from index 0, not including $part1
1349
1350                 $newline = $matches[1];
1351                 $part1 = $matches[2];
1352                 # If the third subpattern matched anything, it will start with |
1353                 if ( $matches[3] !== "" ) {
1354                         $args = explode( "|", substr( $matches[3], 1 ) );
1355                 } else {
1356                         $args = array();
1357                 }
1358                 $argc = count( $args );
1359
1360                 # SUBST
1361                 $mwSubst =& MagicWord::get( MAG_SUBST );
1362                 if ( $mwSubst->matchStartAndRemove( $part1 ) ) {
1363                         if ( $this->mOutputType != OT_WIKI ) {
1364                                 # Invalid SUBST not replaced at PST time
1365                                 # Return without further processing
1366                                 $text = $matches[0];
1367                                 $found = true;
1368                         }
1369                 } elseif ( $this->mOutputType == OT_WIKI ) {
1370                         # SUBST not found in PST pass, do nothing
1371                         $text = $matches[0];
1372                         $found = true;
1373                 }
1374
1375                 # MSG, MSGNW and INT
1376                 if ( !$found ) {
1377                         # Check for MSGNW:
1378                         $mwMsgnw =& MagicWord::get( MAG_MSGNW );
1379                         if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
1380                                 $nowiki = true;
1381                         } else {
1382                                 # Remove obsolete MSG:
1383                                 $mwMsg =& MagicWord::get( MAG_MSG );
1384                                 $mwMsg->matchStartAndRemove( $part1 );
1385                         }
1386
1387                         # Check if it is an internal message
1388                         $mwInt =& MagicWord::get( MAG_INT );
1389                         if ( $mwInt->matchStartAndRemove( $part1 ) ) {
1390                                 if ( $this->incrementIncludeCount( "int:$part1" ) ) {
1391                                         $text = wfMsgReal( $part1, $args, true );
1392                                         $found = true;
1393                                 }
1394                         }
1395                 }
1396
1397                 # NS
1398                 if ( !$found ) {
1399                         # Check for NS: (namespace expansion)
1400                         $mwNs = MagicWord::get( MAG_NS );
1401                         if ( $mwNs->matchStartAndRemove( $part1 ) ) {
1402                                 if ( intval( $part1 ) ) {
1403                                         $text = $wgLang->getNsText( intval( $part1 ) );
1404                                         $found = true;
1405                                 } else {
1406                                         $index = Namespace::getCanonicalIndex( strtolower( $part1 ) );
1407                                         if ( !is_null( $index ) ) {
1408                                                 $text = $wgLang->getNsText( $index );
1409                                                 $found = true;
1410                                         }
1411                                 }
1412                         }
1413                 }
1414
1415                 # LOCALURL and LOCALURLE
1416                 if ( !$found ) {
1417                         $mwLocal = MagicWord::get( MAG_LOCALURL );
1418                         $mwLocalE = MagicWord::get( MAG_LOCALURLE );
1419
1420                         if ( $mwLocal->matchStartAndRemove( $part1 ) ) {
1421                                 $func = 'getLocalURL';
1422                         } elseif ( $mwLocalE->matchStartAndRemove( $part1 ) ) {
1423                                 $func = 'escapeLocalURL';
1424                         } else {
1425                                 $func = '';
1426                         }
1427
1428                         if ( $func !== '' ) {
1429                                 $title = Title::newFromText( $part1 );
1430                                 if ( !is_null( $title ) ) {
1431                                         if ( $argc > 0 ) {
1432                                                 $text = $title->$func( $args[0] );
1433                                         } else {
1434                                                 $text = $title->$func();
1435                                         }
1436                                         $found = true;
1437                                 }
1438                         }
1439                 }
1440
1441                 # Internal variables
1442                 if ( !$found && array_key_exists( $part1, $this->mVariables ) ) {
1443                         $text = $this->mVariables[$part1];
1444                         $found = true;
1445                         $this->mOutput->mContainsOldMagic = true;
1446                 }
1447
1448                 # Arguments input from the caller
1449                 $inputArgs = end( $this->mArgStack );
1450                 if ( !$found && array_key_exists( $part1, $inputArgs ) ) {
1451                         $text = $inputArgs[$part1];
1452                         $found = true;
1453                 }
1454
1455                 # Load from database
1456                 if ( !$found ) {
1457                         $title = Title::newFromText( $part1, NS_TEMPLATE );
1458                         if ( !is_null( $title ) && !$title->isExternal() ) {
1459                                 # Check for excessive inclusion
1460                                 $dbk = $title->getPrefixedDBkey();
1461                                 if ( $this->incrementIncludeCount( $dbk ) ) {
1462                                         $article = new Article( $title );
1463                                         $articleContent = $article->getContentWithoutUsingSoManyDamnGlobals();
1464                                         if ( $articleContent !== false ) {
1465                                                 $found = true;
1466                                                 $text = $articleContent;
1467
1468                                         }
1469                                 }
1470
1471                                 # If the title is valid but undisplayable, make a link to it
1472                                 if ( $this->mOutputType == OT_HTML && !$found ) {
1473                                         $text = "[[" . $title->getPrefixedText() . "]]";
1474                                         $found = true;
1475                                 }
1476                         }
1477                 }
1478
1479                 # Recursive parsing, escaping and link table handling
1480                 # Only for HTML output
1481                 if ( $nowiki && $found && $this->mOutputType == OT_HTML ) {
1482                         $text = wfEscapeWikiText( $text );
1483                 } elseif ( $this->mOutputType == OT_HTML && $found ) {
1484                         # Clean up argument array
1485                         $assocArgs = array();
1486                         $index = 1;
1487                         foreach( $args as $arg ) {
1488                                 $eqpos = strpos( $arg, "=" );
1489                                 if ( $eqpos === false ) {
1490                                         $assocArgs[$index++] = $arg;
1491                                 } else {
1492                                         $name = trim( substr( $arg, 0, $eqpos ) );
1493                                         $value = trim( substr( $arg, $eqpos+1 ) );
1494                                         if ( $value === false ) {
1495                                                 $value = "";
1496                                         }
1497                                         if ( $name !== false ) {
1498                                                 $assocArgs[$name] = $value;
1499                                         }
1500                                 }
1501                         }
1502
1503                         # Do not enter included links in link table
1504                         if ( !is_null( $title ) ) {
1505                                 $wgLinkCache->suspend();
1506                         }
1507
1508                         # Run full parser on the included text
1509                         $text = $this->strip( $text, $this->mStripState );
1510                         $text = $this->internalParse( $text, (bool)$newline, $assocArgs );
1511                         if(!empty($newline)) $text = "\n".$text;
1512
1513                         # Add the result to the strip state for re-inclusion after
1514                         # the rest of the processing
1515                         $text = $this->insertStripItem( $text, $this->mStripState );
1516
1517                         # Resume the link cache and register the inclusion as a link
1518                         if ( !is_null( $title ) ) {
1519                                 $wgLinkCache->resume();
1520                                 $wgLinkCache->addLinkObj( $title );
1521                         }
1522                 }
1523
1524                 if ( !$found ) {
1525                         return $matches[0];
1526                 } else {
1527                         return $text;
1528                 }
1529         }
1530
1531         # Returns true if the function is allowed to include this entity
1532         function incrementIncludeCount( $dbk )
1533         {
1534                 if ( !array_key_exists( $dbk, $this->mIncludeCount ) ) {
1535                         $this->mIncludeCount[$dbk] = 0;
1536                 }
1537                 if ( ++$this->mIncludeCount[$dbk] <= MAX_INCLUDE_REPEAT ) {
1538                         return true;
1539                 } else {
1540                         return false;
1541                 }
1542         }
1543
1544
1545         # Cleans up HTML, removes dangerous tags and attributes
1546         /* private */ function removeHTMLtags( $text )
1547         {
1548                 global $wgUseTidy, $wgUserHtml;
1549                 $fname = "Parser::removeHTMLtags";
1550                 wfProfileIn( $fname );
1551
1552                 if( $wgUserHtml ) {
1553                         $htmlpairs = array( # Tags that must be closed
1554                                 "b", "del", "i", "ins", "u", "font", "big", "small", "sub", "sup", "h1",
1555                                 "h2", "h3", "h4", "h5", "h6", "cite", "code", "em", "s",
1556                                 "strike", "strong", "tt", "var", "div", "center",
1557                                 "blockquote", "ol", "ul", "dl", "table", "caption", "pre",
1558                                 "ruby", "rt" , "rb" , "rp", "p"
1559                         );
1560                         $htmlsingle = array(
1561                                 "br", "hr", "li", "dt", "dd"
1562                         );
1563                         $htmlnest = array( # Tags that can be nested--??
1564                                 "table", "tr", "td", "th", "div", "blockquote", "ol", "ul",
1565                                 "dl", "font", "big", "small", "sub", "sup"
1566                         );
1567                         $tabletags = array( # Can only appear inside table
1568                                 "td", "th", "tr"
1569                         );
1570                 } else {
1571                         $htmlpairs = array();
1572                         $htmlsingle = array();
1573                         $htmlnest = array();
1574                         $tabletags = array();
1575                 }
1576
1577                 $htmlsingle = array_merge( $tabletags, $htmlsingle );
1578                 $htmlelements = array_merge( $htmlsingle, $htmlpairs );
1579
1580                 $htmlattrs = $this->getHTMLattrs () ;
1581
1582                 # Remove HTML comments
1583                 $text = preg_replace( "/(\\n *<!--.*--> *(?=\\n)|<!--.*-->)/sU", "$2", $text );
1584
1585                 $bits = explode( "<", $text );
1586                 $text = array_shift( $bits );
1587                 if(!$wgUseTidy) {
1588                         $tagstack = array(); $tablestack = array();
1589                         foreach ( $bits as $x ) {
1590                                 $prev = error_reporting( E_ALL & ~( E_NOTICE | E_WARNING ) );
1591                                 preg_match( "/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/",
1592                                 $x, $regs );
1593                                 list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
1594                                 error_reporting( $prev );
1595
1596                                 $badtag = 0 ;
1597                                 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
1598                                         # Check our stack
1599                                         if ( $slash ) {
1600                                                 # Closing a tag...
1601                                                 if ( ! in_array( $t, $htmlsingle ) &&
1602                                                 ( count($tagstack) && $ot = array_pop( $tagstack ) ) != $t ) {
1603                                                         if(!empty($ot)) array_push( $tagstack, $ot );
1604                                                         $badtag = 1;
1605                                                 } else {
1606                                                         if ( $t == "table" ) {
1607                                                                 $tagstack = array_pop( $tablestack );
1608                                                         }
1609                                                         $newparams = "";
1610                                                 }
1611                                         } else {
1612                                                 # Keep track for later
1613                                                 if ( in_array( $t, $tabletags ) &&
1614                                                 ! in_array( "table", $tagstack ) ) {
1615                                                         $badtag = 1;
1616                                                 } else if ( in_array( $t, $tagstack ) &&
1617                                                 ! in_array ( $t , $htmlnest ) ) {
1618                                                         $badtag = 1 ;
1619                                                 } else if ( ! in_array( $t, $htmlsingle ) ) {
1620                                                         if ( $t == "table" ) {
1621                                                                 array_push( $tablestack, $tagstack );
1622                                                                 $tagstack = array();
1623                                                         }
1624                                                         array_push( $tagstack, $t );
1625                                                 }
1626                                                 # Strip non-approved attributes from the tag
1627                                                 $newparams = $this->fixTagAttributes($params);
1628
1629                                         }
1630                                         if ( ! $badtag ) {
1631                                                 $rest = str_replace( ">", "&gt;", $rest );
1632                                                 $text .= "<$slash$t $newparams$brace$rest";
1633                                                 continue;
1634                                         }
1635                                 }
1636                                 $text .= "&lt;" . str_replace( ">", "&gt;", $x);
1637                         }
1638                         # Close off any remaining tags
1639                         while ( $t = array_pop( $tagstack ) ) {
1640                                 $text .= "</$t>\n";
1641                                 if ( $t == "table" ) { $tagstack = array_pop( $tablestack ); }
1642                         }
1643                 } else {
1644                         # this might be possible using tidy itself
1645                         foreach ( $bits as $x ) {
1646                                 preg_match( "/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/",
1647                                 $x, $regs );
1648                                 @list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
1649                                 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
1650                                         $newparams = $this->fixTagAttributes($params);
1651                                         $rest = str_replace( ">", "&gt;", $rest );
1652                                         $text .= "<$slash$t $newparams$brace$rest";
1653                                 } else {
1654                                         $text .= "&lt;" . str_replace( ">", "&gt;", $x);
1655                                 }
1656                         }
1657                 }
1658                 wfProfileOut( $fname );
1659                 return $text;
1660         }
1661
1662
1663 /*
1664  *
1665  * This function accomplishes several tasks:
1666  * 1) Auto-number headings if that option is enabled
1667  * 2) Add an [edit] link to sections for logged in users who have enabled the option
1668  * 3) Add a Table of contents on the top for users who have enabled the option
1669  * 4) Auto-anchor headings
1670  *
1671  * It loops through all headlines, collects the necessary data, then splits up the
1672  * string and re-inserts the newly formatted headlines.
1673  *
1674  */
1675
1676         /* private */ function formatHeadings( $text )
1677         {
1678                 global $wgInputEncoding;
1679
1680                 $doNumberHeadings = $this->mOptions->getNumberHeadings();
1681                 $doShowToc = $this->mOptions->getShowToc();
1682                 if( !$this->mTitle->userCanEdit() ) {
1683                         $showEditLink = 0;
1684                         $rightClickHack = 0;
1685                 } else {
1686                         $showEditLink = $this->mOptions->getEditSection();
1687                         $rightClickHack = $this->mOptions->getEditSectionOnRightClick();
1688                 }
1689
1690                 # Inhibit editsection links if requested in the page
1691                 $esw =& MagicWord::get( MAG_NOEDITSECTION );
1692                 if( $esw->matchAndRemove( $text ) ) {
1693                         $showEditLink = 0;
1694                 }
1695                 # if the string __NOTOC__ (not case-sensitive) occurs in the HTML,
1696                 # do not add TOC
1697                 $mw =& MagicWord::get( MAG_NOTOC );
1698                 if( $mw->matchAndRemove( $text ) ) {
1699                         $doShowToc = 0;
1700                 }
1701
1702                 # never add the TOC to the Main Page. This is an entry page that should not
1703                 # be more than 1-2 screens large anyway
1704                 if( $this->mTitle->getPrefixedText() == wfMsg("mainpage") ) {
1705                         $doShowToc = 0;
1706                 }
1707
1708                 # Get all headlines for numbering them and adding funky stuff like [edit]
1709                 # links - this is for later, but we need the number of headlines right now
1710                 $numMatches = preg_match_all( "/<H([1-6])(.*?" . ">)(.*?)<\/H[1-6]>/i", $text, $matches );
1711
1712                 # if there are fewer than 4 headlines in the article, do not show TOC
1713                 if( $numMatches < 4 ) {
1714                         $doShowToc = 0;
1715                 }
1716
1717                 # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
1718                 # override above conditions and always show TOC
1719                 $mw =& MagicWord::get( MAG_FORCETOC );
1720                 if ($mw->matchAndRemove( $text ) ) {
1721                         $doShowToc = 1;
1722                 }
1723
1724
1725                 # We need this to perform operations on the HTML
1726                 $sk =& $this->mOptions->getSkin();
1727
1728                 # headline counter
1729                 $headlineCount = 0;
1730
1731                 # Ugh .. the TOC should have neat indentation levels which can be
1732                 # passed to the skin functions. These are determined here
1733                 $toclevel = 0;
1734                 $toc = "";
1735                 $full = "";
1736                 $head = array();
1737                 $sublevelCount = array();
1738                 $level = 0;
1739                 $prevlevel = 0;
1740                 foreach( $matches[3] as $headline ) {
1741                         $numbering = "";
1742                         if( $level ) {
1743                                 $prevlevel = $level;
1744                         }
1745                         $level = $matches[1][$headlineCount];
1746                         if( ( $doNumberHeadings || $doShowToc ) && $prevlevel && $level > $prevlevel ) {
1747                                 # reset when we enter a new level
1748                                 $sublevelCount[$level] = 0;
1749                                 $toc .= $sk->tocIndent( $level - $prevlevel );
1750                                 $toclevel += $level - $prevlevel;
1751                         }
1752                         if( ( $doNumberHeadings || $doShowToc ) && $level < $prevlevel ) {
1753                                 # reset when we step back a level
1754                                 $sublevelCount[$level+1]=0;
1755                                 $toc .= $sk->tocUnindent( $prevlevel - $level );
1756                                 $toclevel -= $prevlevel - $level;
1757                         }
1758                         # count number of headlines for each level
1759                         @$sublevelCount[$level]++;
1760                         if( $doNumberHeadings || $doShowToc ) {
1761                                 $dot = 0;
1762                                 for( $i = 1; $i <= $level; $i++ ) {
1763                                         if( !empty( $sublevelCount[$i] ) ) {
1764                                                 if( $dot ) {
1765                                                         $numbering .= ".";
1766                                                 }
1767                                                 $numbering .= $sublevelCount[$i];
1768                                                 $dot = 1;
1769                                         }
1770                                 }
1771                         }
1772
1773                         # The canonized header is a version of the header text safe to use for links
1774                         # Avoid insertion of weird stuff like <math> by expanding the relevant sections
1775                         $canonized_headline = $this->unstrip( $headline, $this->mStripState );
1776
1777                         # strip out HTML
1778                         $canonized_headline = preg_replace( "/<.*?" . ">/","",$canonized_headline );
1779                         $tocline = trim( $canonized_headline );
1780                         $canonized_headline = preg_replace("/[ \\?&\\/<>\\(\\)\\[\\]=,+']+/", '_', urlencode( do_html_entity_decode( $tocline, ENT_COMPAT, $wgInputEncoding ) ) );
1781                         $refer[$headlineCount] = $canonized_headline;
1782
1783                         # count how many in assoc. array so we can track dupes in anchors
1784                         @$refers[$canonized_headline]++;
1785                         $refcount[$headlineCount]=$refers[$canonized_headline];
1786
1787                         # Prepend the number to the heading text
1788
1789                         if( $doNumberHeadings || $doShowToc ) {
1790                                 $tocline = $numbering . " " . $tocline;
1791
1792                                 # Don't number the heading if it is the only one (looks silly)
1793                                 if( $doNumberHeadings && count( $matches[3] ) > 1) {
1794                                         # the two are different if the line contains a link
1795                                         $headline=$numbering . " " . $headline;
1796                                 }
1797                         }
1798
1799                         # Create the anchor for linking from the TOC to the section
1800                         $anchor = $canonized_headline;
1801                         if($refcount[$headlineCount] > 1 ) {
1802                                 $anchor .= "_" . $refcount[$headlineCount];
1803                         }
1804                         if( $doShowToc ) {
1805                                 $toc .= $sk->tocLine($anchor,$tocline,$toclevel);
1806                         }
1807                         if( $showEditLink ) {
1808                                 if ( empty( $head[$headlineCount] ) ) {
1809                                         $head[$headlineCount] = "";
1810                                 }
1811                                 $head[$headlineCount] .= $sk->editSectionLink($headlineCount+1);
1812                         }
1813
1814                         # Add the edit section span
1815                         if( $rightClickHack ) {
1816                                 $headline = $sk->editSectionScript($headlineCount+1,$headline);
1817                         }
1818
1819                         # give headline the correct <h#> tag
1820                         @$head[$headlineCount] .= "<a name=\"$anchor\"></a><h".$level.$matches[2][$headlineCount] .$headline."</h".$level.">";
1821
1822                         $headlineCount++;
1823                 }
1824
1825                 if( $doShowToc ) {
1826                         $toclines = $headlineCount;
1827                         $toc .= $sk->tocUnindent( $toclevel );
1828                         $toc = $sk->tocTable( $toc );
1829                 }
1830
1831                 # split up and insert constructed headlines
1832
1833                 $blocks = preg_split( "/<H[1-6].*?" . ">.*?<\/H[1-6]>/i", $text );
1834                 $i = 0;
1835
1836                 foreach( $blocks as $block ) {
1837                         if( $showEditLink && $headlineCount > 0 && $i == 0 && $block != "\n" ) {
1838                             # This is the [edit] link that appears for the top block of text when
1839                                 # section editing is enabled
1840
1841                                 # Disabled because it broke block formatting
1842                                 # For example, a bullet point in the top line
1843                                 # $full .= $sk->editSectionLink(0);
1844                         }
1845                         $full .= $block;
1846                         if( $doShowToc && !$i) {
1847                         # Top anchor now in skin
1848                                 $full = $full.$toc;
1849                         }
1850
1851                         if( !empty( $head[$i] ) ) {
1852                                 $full .= $head[$i];
1853                         }
1854                         $i++;
1855                 }
1856
1857                 return $full;
1858         }
1859
1860         /* private */ function doMagicISBN( &$tokenizer )
1861         {
1862                 global $wgLang;
1863
1864                 # Check whether next token is a text token
1865                 # If yes, fetch it and convert the text into a
1866                 # Special::BookSources link
1867                 $token = $tokenizer->previewToken();
1868                 while ( $token["type"] == "" )
1869                 {
1870                         $tokenizer->nextToken();
1871                         $token = $tokenizer->previewToken();
1872                 }
1873                 if ( $token["type"] == "text" )
1874                 {
1875                         $token = $tokenizer->nextToken();
1876                         $x = $token["text"];
1877                         $valid = "0123456789-ABCDEFGHIJKLMNOPQRSTUVWXYZ";
1878
1879                         $isbn = $blank = "" ;
1880                         while ( " " == $x{0} ) {
1881                                 $blank .= " ";
1882                                 $x = substr( $x, 1 );
1883                         }
1884                         while ( strstr( $valid, $x{0} ) != false ) {
1885                                 $isbn .= $x{0};
1886                                 $x = substr( $x, 1 );
1887                         }
1888                         $num = str_replace( "-", "", $isbn );
1889                         $num = str_replace( " ", "", $num );
1890
1891                         if ( "" == $num ) {
1892                                 $text = "ISBN $blank$x";
1893                         } else {
1894                                 $titleObj = Title::makeTitle( NS_SPECIAL, "Booksources" );
1895                                 $text = "<a href=\"" .
1896                                 $titleObj->escapeLocalUrl( "isbn={$num}" ) .
1897                                         "\" class=\"internal\">ISBN $isbn</a>";
1898                                 $text .= $x;
1899                         }
1900                 } else {
1901                         $text = "ISBN ";
1902                 }
1903                 return $text;
1904         }
1905         /* private */ function doMagicRFC( &$tokenizer )
1906         {
1907                 global $wgLang;
1908
1909                 # Check whether next token is a text token
1910                 # If yes, fetch it and convert the text into a
1911                 # link to an RFC source
1912                 $token = $tokenizer->previewToken();
1913                 while ( $token["type"] == "" )
1914                 {
1915                         $tokenizer->nextToken();
1916                         $token = $tokenizer->previewToken();
1917                 }
1918                 if ( $token["type"] == "text" )
1919                 {
1920                         $token = $tokenizer->nextToken();
1921                         $x = $token["text"];
1922                         $valid = "0123456789";
1923
1924                         $rfc = $blank = "" ;
1925                         while ( " " == $x{0} ) {
1926                                 $blank .= " ";
1927                                 $x = substr( $x, 1 );
1928                         }
1929                         while ( strstr( $valid, $x{0} ) != false ) {
1930                                 $rfc .= $x{0};
1931                                 $x = substr( $x, 1 );
1932                         }
1933
1934                         if ( "" == $rfc ) {
1935                                 $text .= "RFC $blank$x";
1936                         } else {
1937                                 $url = wfmsg( "rfcurl" );
1938                                 $url = str_replace( "$1", $rfc, $url);
1939                                 $sk =& $this->mOptions->getSkin();
1940                                 $la = $sk->getExternalLinkAttributes( $url, "RFC {$rfc}" );
1941                                 $text = "<a href='{$url}'{$la}>RFC {$rfc}</a>{$x}";
1942                         }
1943                 } else {
1944                         $text = "RFC ";
1945                 }
1946                 return $text;
1947         }
1948
1949         function preSaveTransform( $text, &$title, &$user, $options, $clearState = true )
1950         {
1951                 $this->mOptions = $options;
1952                 $this->mTitle =& $title;
1953                 $this->mOutputType = OT_WIKI;
1954
1955                 if ( $clearState ) {
1956                         $this->clearState();
1957                 }
1958
1959                 $stripState = false;
1960                 $pairs = array(
1961                         "\r\n" => "\n",
1962                         );
1963                 $text = str_replace(array_keys($pairs), array_values($pairs), $text);
1964                 // now with regexes
1965                 $pairs = array(
1966                         "/<br.+(clear|break)=[\"']?(all|both)[\"']?\\/?>/i" => '<br style="clear:both;"/>',
1967                         "/<br *?>/i" => "<br/>",
1968                 );
1969                 $text = preg_replace(array_keys($pairs), array_values($pairs), $text);
1970                 $text = $this->strip( $text, $stripState, false );
1971                 $text = $this->pstPass2( $text, $user );
1972                 $text = $this->unstrip( $text, $stripState );
1973                 return $text;
1974         }
1975
1976         /* private */ function pstPass2( $text, &$user )
1977         {
1978                 global $wgLang, $wgLocaltimezone, $wgCurParser;
1979
1980                 # Variable replacement
1981                 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
1982                 $text = $this->replaceVariables( $text );
1983
1984                 # Signatures
1985                 #
1986                 $n = $user->getName();
1987                 $k = $user->getOption( "nickname" );
1988                 if ( "" == $k ) { $k = $n; }
1989                 if(isset($wgLocaltimezone)) {
1990                         $oldtz = getenv("TZ"); putenv("TZ=$wgLocaltimezone");
1991                 }
1992                 /* Note: this is an ugly timezone hack for the European wikis */
1993                 $d = $wgLang->timeanddate( date( "YmdHis" ), false ) .
1994                   " (" . date( "T" ) . ")";
1995                 if(isset($wgLocaltimezone)) putenv("TZ=$oldtz");
1996
1997                 $text = preg_replace( "/~~~~~/", $d, $text );
1998                 $text = preg_replace( "/~~~~/", "[[" . $wgLang->getNsText(
1999                   Namespace::getUser() ) . ":$n|$k]] $d", $text );
2000                 $text = preg_replace( "/~~~/", "[[" . $wgLang->getNsText(
2001                   Namespace::getUser() ) . ":$n|$k]]", $text );
2002
2003                 # Context links: [[|name]] and [[name (context)|]]
2004                 #
2005                 $tc = "[&;%\\-,.\\(\\)' _0-9A-Za-z\\/:\\x80-\\xff]";
2006                 $np = "[&;%\\-,.' _0-9A-Za-z\\/:\\x80-\\xff]"; # No parens
2007                 $namespacechar = '[ _0-9A-Za-z\x80-\xff]'; # Namespaces can use non-ascii!
2008                 $conpat = "/^({$np}+) \\(({$tc}+)\\)$/";
2009
2010                 $p1 = "/\[\[({$np}+) \\(({$np}+)\\)\\|]]/";             # [[page (context)|]]
2011                 $p2 = "/\[\[\\|({$tc}+)]]/";                                    # [[|page]]
2012                 $p3 = "/\[\[($namespacechar+):({$np}+)\\|]]/";          # [[namespace:page|]]
2013                 $p4 = "/\[\[($namespacechar+):({$np}+) \\(({$np}+)\\)\\|]]/";
2014                                                                                                                 # [[ns:page (cont)|]]
2015                 $context = "";
2016                 $t = $this->mTitle->getText();
2017                 if ( preg_match( $conpat, $t, $m ) ) {
2018                         $context = $m[2];
2019                 }
2020                 $text = preg_replace( $p4, "[[\\1:\\2 (\\3)|\\2]]", $text );
2021                 $text = preg_replace( $p1, "[[\\1 (\\2)|\\1]]", $text );
2022                 $text = preg_replace( $p3, "[[\\1:\\2|\\2]]", $text );
2023
2024                 if ( "" == $context ) {
2025                         $text = preg_replace( $p2, "[[\\1]]", $text );
2026                 } else {
2027                         $text = preg_replace( $p2, "[[\\1 ({$context})|\\1]]", $text );
2028                 }
2029
2030                 /*
2031                 $mw =& MagicWord::get( MAG_SUBST );
2032                 $wgCurParser = $this->fork();
2033                 $text = $mw->substituteCallback( $text, "wfBraceSubstitution" );
2034                 $this->merge( $wgCurParser );
2035                 */
2036
2037                 # Trim trailing whitespace
2038                 # MAG_END (__END__) tag allows for trailing
2039                 # whitespace to be deliberately included
2040                 $text = rtrim( $text );
2041                 $mw =& MagicWord::get( MAG_END );
2042                 $mw->matchAndRemove( $text );
2043
2044                 return $text;
2045         }
2046
2047         # Set up some variables which are usually set up in parse()
2048         # so that an external function can call some class members with confidence
2049         function startExternalParse( &$title, $options, $outputType, $clearState = true )
2050         {
2051                 $this->mTitle =& $title;
2052                 $this->mOptions = $options;
2053                 $this->mOutputType = $outputType;
2054                 if ( $clearState ) {
2055                         $this->clearState();
2056                 }
2057         }
2058
2059         function transformMsg( $text, $options ) {
2060                 global $wgTitle;
2061                 static $executing = false;
2062
2063                 # Guard against infinite recursion
2064                 if ( $executing ) {
2065                         return $text;
2066                 }
2067                 $executing = true;
2068
2069                 $this->mTitle = $wgTitle;
2070                 $this->mOptions = $options;
2071                 $this->mOutputType = OT_MSG;
2072                 $this->clearState();
2073                 $text = $this->replaceVariables( $text );
2074
2075                 $executing = false;
2076                 return $text;
2077         }
2078 }
2079
2080 class ParserOutput
2081 {
2082         var $mText, $mLanguageLinks, $mCategoryLinks, $mContainsOldMagic;
2083
2084         function ParserOutput( $text = "", $languageLinks = array(), $categoryLinks = array(),
2085                 $containsOldMagic = false )
2086         {
2087                 $this->mText = $text;
2088                 $this->mLanguageLinks = $languageLinks;
2089                 $this->mCategoryLinks = $categoryLinks;
2090                 $this->mContainsOldMagic = $containsOldMagic;
2091         }
2092
2093         function getText() { return $this->mText; }
2094         function getLanguageLinks() { return $this->mLanguageLinks; }
2095         function getCategoryLinks() { return $this->mCategoryLinks; }
2096         function containsOldMagic() { return $this->mContainsOldMagic; }
2097         function setText( $text ) { return wfSetVar( $this->mText, $text ); }
2098         function setLanguageLinks( $ll ) { return wfSetVar( $this->mLanguageLinks, $ll ); }
2099         function setCategoryLinks( $cl ) { return wfSetVar( $this->mCategoryLinks, $cl ); }
2100         function setContainsOldMagic( $com ) { return wfSetVar( $this->mContainsOldMagic, $com ); }
2101
2102         function merge( $other ) {
2103                 $this->mLanguageLinks = array_merge( $this->mLanguageLinks, $other->mLanguageLinks );
2104                 $this->mCategoryLinks = array_merge( $this->mCategoryLinks, $this->mLanguageLinks );
2105                 $this->mContainsOldMagic = $this->mContainsOldMagic || $other->mContainsOldMagic;
2106         }
2107
2108 }
2109
2110 class ParserOptions
2111 {
2112         # All variables are private
2113         var $mUseTeX;                    # Use texvc to expand <math> tags
2114         var $mUseCategoryMagic;          # Treat [[Category:xxxx]] tags specially
2115         var $mUseDynamicDates;           # Use $wgDateFormatter to format dates
2116         var $mInterwikiMagic;            # Interlanguage links are removed and returned in an array
2117         var $mAllowExternalImages;       # Allow external images inline
2118         var $mSkin;                      # Reference to the preferred skin
2119         var $mDateFormat;                # Date format index
2120         var $mEditSection;               # Create "edit section" links
2121         var $mEditSectionOnRightClick;   # Generate JavaScript to edit section on right click
2122         var $mNumberHeadings;            # Automatically number headings
2123         var $mShowToc;                   # Show table of contents
2124
2125         function getUseTeX() { return $this->mUseTeX; }
2126         function getUseCategoryMagic() { return $this->mUseCategoryMagic; }
2127         function getUseDynamicDates() { return $this->mUseDynamicDates; }
2128         function getInterwikiMagic() { return $this->mInterwikiMagic; }
2129         function getAllowExternalImages() { return $this->mAllowExternalImages; }
2130         function getSkin() { return $this->mSkin; }
2131         function getDateFormat() { return $this->mDateFormat; }
2132         function getEditSection() { return $this->mEditSection; }
2133         function getEditSectionOnRightClick() { return $this->mEditSectionOnRightClick; }
2134         function getNumberHeadings() { return $this->mNumberHeadings; }
2135         function getShowToc() { return $this->mShowToc; }
2136
2137         function setUseTeX( $x ) { return wfSetVar( $this->mUseTeX, $x ); }
2138         function setUseCategoryMagic( $x ) { return wfSetVar( $this->mUseCategoryMagic, $x ); }
2139         function setUseDynamicDates( $x ) { return wfSetVar( $this->mUseDynamicDates, $x ); }
2140         function setInterwikiMagic( $x ) { return wfSetVar( $this->mInterwikiMagic, $x ); }
2141         function setAllowExternalImages( $x ) { return wfSetVar( $this->mAllowExternalImages, $x ); }
2142         function setSkin( $x ) { return wfSetRef( $this->mSkin, $x ); }
2143         function setDateFormat( $x ) { return wfSetVar( $this->mDateFormat, $x ); }
2144         function setEditSection( $x ) { return wfSetVar( $this->mEditSection, $x ); }
2145         function setEditSectionOnRightClick( $x ) { return wfSetVar( $this->mEditSectionOnRightClick, $x ); }
2146         function setNumberHeadings( $x ) { return wfSetVar( $this->mNumberHeadings, $x ); }
2147         function setShowToc( $x ) { return wfSetVar( $this->mShowToc, $x ); }
2148
2149         /* static */ function newFromUser( &$user )
2150         {
2151                 $popts = new ParserOptions;
2152                 $popts->initialiseFromUser( $user );
2153                 return $popts;
2154         }
2155
2156         function initialiseFromUser( &$userInput )
2157         {
2158                 global $wgUseTeX, $wgUseCategoryMagic, $wgUseDynamicDates, $wgInterwikiMagic, $wgAllowExternalImages;
2159
2160                 if ( !$userInput ) {
2161                         $user = new User;
2162                         $user->setLoaded( true );
2163                 } else {
2164                         $user =& $userInput;
2165                 }
2166
2167                 $this->mUseTeX = $wgUseTeX;
2168                 $this->mUseCategoryMagic = $wgUseCategoryMagic;
2169                 $this->mUseDynamicDates = $wgUseDynamicDates;
2170                 $this->mInterwikiMagic = $wgInterwikiMagic;
2171                 $this->mAllowExternalImages = $wgAllowExternalImages;
2172                 $this->mSkin =& $user->getSkin();
2173                 $this->mDateFormat = $user->getOption( "date" );
2174                 $this->mEditSection = $user->getOption( "editsection" );
2175                 $this->mEditSectionOnRightClick = $user->getOption( "editsectiononrightclick" );
2176                 $this->mNumberHeadings = $user->getOption( "numberheadings" );
2177                 $this->mShowToc = $user->getOption( "showtoc" );
2178         }
2179
2180
2181 }
2182
2183 # Regex callbacks, used in Parser::replaceVariables
2184 function wfBraceSubstitution( $matches )
2185 {
2186         global $wgCurParser;
2187         return $wgCurParser->braceSubstitution( $matches );
2188 }
2189
2190 ?>